The functionality of alarm sources has to be proved on a product and furthermore has to be constantly checked in the field.
Conventional solutions implement tests for alarms in hardware and/or in software. Disadvantages of these implementations are that the tests can be performed only after the chip is completely running and that special circuits are necessary for testing the functionality of the alarms in the different parts of the chips. As an example, there has to be an extra signal which performs a test and a signal which marks during the test that a test is being performed and that the system has to ignore the generated alarm. In this context, it is important to mention that in the case of an alarm often not only an alarm state is signaled but also the generating module activates further mechanisms which ensure that the chip is no longer functional and an alarm has occurred. As an example, a CPU (Central Processing Unit) not only signals the alarm state but also stops a performing of instructions. Such redundancy measures have to be taken separately in the case of a test of the alarm source to ensure that after the test the system can continue to be used.